OSX install matplotlib without anaconda

If you ever tried to install matplotlib without using anaconda, then you probably pulled out some hair. After destroying my python environment several times, I finally figured it out. This solution only works for Mac, so sorry Linux and Windows users.

I recommend using pipenv to manage python version, because downloading python from brew was a pain in the ass when you need more than one version of python. I needed to use python 3.6, because at the time tensorflow did not support python 3.7.

First update brew and install pyenv pipenv

$ brew update
$ brew install pyenv pipenv

Create project directory

$ mkdir ~/project
$ cd ~/project
$ pyenv init

Install desired python version.

$ PYTHON_CONFIGURE_OPTS="--enable-framework" CFLAGS="-I$(brew --prefix openssl)/include" LDFLAGS="-L$(brew --prefix openssl)/lib" pyenv install 3.6.5

Access installed version of python

$ pyenv local 3.6.5
$ pipenv shell
$ python --version

Install dependencies

$ pip3 install numpy
$ pip3 install matplotlib

Create file example.py and add the following in it:

import numpy as np
import matplotlib.pyplot as plt

x_data = np.linspace(0, 10, 10) + np.random.uniform(-1.5, 1.5, 10)
y_data = np.linspace(0, 10, 10) + np.random.uniform(-1.5, 1.5, 10)
plt.plot(x_data, y_data)
plt.show()

Run script

$ python3 example.py

Voila.